> May I ask what in particular gives you confidence that these two companies know what they are doing wrt cybersecurity.

It's not that I think OpenAI and Anthropic know what they're doing - although I can see why my slightly flippant comment might make you think that - but that I just don't trust anyone with these agents, especially when the whole point of using them would be to do that with such critical data.